IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
For a vectored read with payload count>1 will make two read
requests and to hold that a single contiguous memory is allocated.
So after completing the read request, instead of sending as vector
we will aggregate all the reads one.
Change-Id: I15e7d7bddc1a62d5097a39392575f47cfff3d3a8
BUG: 1164079
Signed-off-by: Mohammed Rafi KC <rkavunga@redhat.com>
Reviewed-on: http://review.gluster.org/9321
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Tested-by: Raghavendra G <rgowdapp@redhat.com>
In the patch http://review.gluster.org/#/c/9488/ the default
configuration for server and client protocol xlators, had
event-threads set to 2, but the default conf was not updated
with these values, hence the default threads remained at 1.
This patch corrects the same by first updating the thread count
from the default configuration before updating it with the
latest.
Change-Id: I3ce54053a59ca796b004fa5462e43ca19a5f2915
BUG: 1104462
Signed-off-by: Shyam <srangana@redhat.com>
Reviewed-on: http://review.gluster.org/9604
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
... from all bricks in the volume
This patch is important in the context of MT epoll. With MT epoll,
notification events from client xlators could reach cluster xlators like
afr, dht, ec, stripe etc. in different orders.
For e.g, In a distributed replicate volume of 2 bricks, namely Brick1
and Brick2, the following network events are observed by a mount
process.
- connection to Brick1 is broken.
- connection to Brick1 has been restored.
- connection to Brick2 is broken.
- connection to Brick2 has been restored.
Without establishing a total ordering of events, we can't guarantee that
cluster xlators like afr, dht perceive them in the same order. While we
would expect afr (say) to perceive it as only one of Brick1 and Brick2
going down at any given time, it is possible for the notification of
Brick2 going offline to race with the notification of Brick1 coming back
online.
Change-Id: I78f5a52bfb05593335d0e9ad53ebfff98995593d
BUG: 1104462
Signed-off-by: Krishnan Parthasarathi <kparthas@redhat.com>
Reviewed-on: http://review.gluster.org/9591
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Add the ability to configure the number of event threads
for various gluster services.
Currently with the multi thread epoll patch, it is possible
to have more than one thread waiting on socket activity and
processing the same. This thread count is currently static,
which this commit makes dynamic.
The current services which use IO path, i.e brick processes,
any client process (nfs, FUSE, gfapi, heal,
rebalance, etc.a), gain 2 set parameters to control the number
of threads that are processing events. These settings are,
- client.event-threads <n>
- server.event-threads <n>
The client setting affects the client graph consumers, and the
server setting affects the brick processes. These are processed
and inited/reconfigured using the client/server protocol xlators.
Other services (say glusterd) would need to extend similar
configuration settings to take advantage of multi threaded event
processing.
At present glusterd is not enabled with this commit, as it does not
stand to gain from this multi-threading (as I understand it).
Change-Id: Id8422fc57a9f95a135158eb6477ccf9d3c9ea4d9
BUG: 1104462
Signed-off-by: Shyam <srangana@redhat.com>
Reviewed-on: http://review.gluster.org/9488
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
- edge triggered (oneshot) polling with epoll
- pick one event to avoid multiple events getting picked up by same
thread
and so get better distribution of events against multiple threads
- wire support for multiple poll threads to epoll_wait in parallel
- evdata to store absolute index and not hint for epoll
- store index and gen of slot instead of fd and index hint
- perform fd close asynchronously inside event.c for multithread safety
- poll is still single threaded
Change-Id: I536851dda0ab224c5d5a1b130a571397c9cace8f
BUG: 1104462
Signed-off-by: Anand Avati <avati@redhat.com>
Signed-off-by: Vijaikumar M <vmallika@redhat.com>
Signed-off-by: Jeff Darcy <jdarcy@redhat.com>
Signed-off-by: Shyam <srangana@redhat.com>
Reviewed-on: http://review.gluster.org/3842
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
ftw provides file tree walk.
dir_scan does just a readdir not readdirp.
Also changed Afr's self-heal-daemon's crawling functions to use this.
These utils will be used by ec in future to do proactive/full healing.
Change-Id: I05715ddb789592c1b79a71e98f1e8cc29aac5c26
BUG: 1177601
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/9485
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Ravishankar N <ravishankar@redhat.com>
Reviewed-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Anonymous file descriptors need to be handled specially because
they can be used in some non standard ways (i.e. an anonymous fd
can be used without having been opened).
This caused NFS to fail on some operations because ec always
expected to have a previous successful opendir call (from patch
http://review.gluster.org/9098/).
This patch treats all anonymous fd as opened on all subvolumes.
Change-Id: I09dbbce2ffc1ae3a5bcbb328bed55b84f4f0b9f8
BUG: 1187474
Signed-off-by: Xavier Hernandez <xhernandez@datalab.es>
Reviewed-on: http://review.gluster.org/9513
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
PROBLEM:
Previously gluster accepting input value as a percentage which is out of range
[0-100] and accepting input value as a size (unit is byte) which is fractional
for option cluster.min-free-disk.
FIX:
Now with this change it will refer to correct validation function
and it will accept value that is in range [0-100] for input value as a
percentage and unsigned integer value for input as a size (unit in byte)
for option cluster.min-free-disk.
Change-Id: Iee1962a100542e146276cfc8a4068abddee2bf2d
BUG: 1163108
Signed-off-by: Gaurav Kumar Garg <ggarg@redhat.com>
Reviewed-on: http://review.gluster.org/9104
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
When lookup has succeeded on multiple subvols of AFR (including
the read child of the parent dir) and all of them are "readable",
ideally the call must be unwound with postparent from the parent's
read child. But that is not the case, due to a bug introduced in
the commit c78998c39f0857ea7aacba360632c148afc54a55. This patch
fixes the issue.
Change-Id: I83b0c26494a5d0bdbc30fcbe974fbdb6f7e9c84a
BUG: 1179169
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/9569
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Problem : heald.t uses EXPECT to check whether shd process is up or not, but as
shd is spawned with NO_WAIT end of volume start transaction doesn't gurantee
that the process will be up by that time.
Solution : Use EXPECT_WITHIN instead of EXPECT
Change-Id: Ic81725aa7e7cde9c0c873837fcc4a73d8318dfa0
BUG: 1163543
Signed-off-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-on: http://review.gluster.org/9575
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Reviewed-by: Shyamsundar Ranganathan <srangana@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>
Problem : There was mismatch between arguments and parameters in some functions
(ex:glusterfs_uuid_buf_get,glusterfs_lkowner_buf_get). It could lead to stack overflow issues .
Fix : Arguments are removed during calling these function.
Change-Id: Icb41bd4119502d192d9cc7242d385ebe62cdb51a
BUG: 1180424
Signed-off-by: Anand <anekkunt@redhat.com>
Reviewed-on: http://review.gluster.org/9427
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Gaurav Kumar Garg <ggarg@redhat.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Checks for rpmbuild/mock is removed from run-tests since this check
is not in the glusterfs-regression-tests package.
Alongside, the dependency mock is also removed from glusterfs.spec.in
as mock is only used by basic/rpm.t.
Change-Id: I5a9036141d0900b3a0bedd95b8c58fb0cdc13ca1
BUG: 1178008
Signed-off-by: Nandaja Varma <nvarma@redhat.com>
Reviewed-on: http://review.gluster.org/9512
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>
Reviewed-by: Lalatendu Mohanty <lmohanty@redhat.com>
Reviewed-by: Humble Devassy Chirammal <humble.devassy@gmail.com>
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
This is to prevent spurious heals that can result in self-heal.
Change-Id: I0b27c1c1fc7a58e2683cb1ca135117a85efcc6c9
BUG: 1179180
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/9523
Reviewed-by: Xavier Hernandez <xhernandez@datalab.es>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Tested-by: Vijay Bellur <vbellur@redhat.com>
Problem:
"make dist" gives the following error:
tar: bug-1140162-file-snapshot-and-features-encryption-option-validation.t:
file name is too long (max 99); not dumped
Here, .t file is not getting included on the "make dist" tarball as file
name is too long for the "tar" command. Result is, upon distributing the tests
through the tarball, this particular test will not get run on the target system.
Solution:
Rename the file to a shorter one(less than 99) to avoid this error.
Change-Id: I29c8da649b8b5e00b3b4dada02c8b69b2d7f0e2c
BUG: 1140162
Signed-off-by: Saravanakumar Arumugam <sarumuga@redhat.com>
Reviewed-on: http://review.gluster.org/9511
Reviewed-by: Gaurav Kumar Garg <ggarg@redhat.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>
Reviewed-by: Sachin Pandit <spandit@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Tested-by: Vijay Bellur <vbellur@redhat.com>
That way a lookup would be forced on the entry, and its attributes will
always be selected from its read subvol.
Change-Id: Iaba25e2cd5f83e983fc8b1a1f48da3850808e6b8
BUG: 1179169
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/9477
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Recent modification in sparse-file-self-heal.t added a
truncate -s 1G, and that unit was not emulated correctly
for non Linux systems. As a result, all regression tests
hang. Fill the gap to restore regression.
BUG: 1129939
Change-Id: Ib45376b4b2e74d1868f3ebdd5564b2186b4318fa
Signed-off-by: Emmanuel Dreyfus <manu@netbsd.org>
Reviewed-on: http://review.gluster.org/9519
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Remove the caches from the 'dist-hook' target that 'make dist' executes.
Also, there is no need to run ./configure inside contrib/argp-standalone
if it will not get used. configure.ac has been modified to only
configure the contrib/argp-standalone sources for distributions that do
not have the <argp.h> header available.
BUG: 1183538
Change-Id: Iedd67139ec83f0076e24ed0d575f9053ac1a94ec
Signed-off-by: Niels de Vos <ndevos@redhat.com>
Reviewed-on: http://review.gluster.org/9463
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Kaleb KEITHLEY <kkeithle@redhat.com>
Problem:
When data-self-heal-algorithm is set to 'full', shd just reads from
source and writes to sink. If source file happened to be sparse (VM
workloads), we end up actually writing 0s to the corresponding regions
of the sink causing it to lose its sparseness.
Fix:
If the source file is sparse, and the data read from source and sink are
both zeros for that range, skip writing that range to the sink.
Change-Id: I787b06a553803247f43a40c00139cb483a22f9ca
BUG: 1166020
Signed-off-by: Ravishankar N <ravishankar@redhat.com>
Reviewed-on: http://review.gluster.org/9480
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
In a n-way replication, where n>=3 fail snapshot,
even if one brick is down.
Also check for glusterd quorum, irrespective of the force option
Modified testcase tests/bugs/snapshot/bug-1090042.t because
it tested the successful creation of snapshot with force
command.
Change-Id: I72666f8f1484bd1766b9d6799c20766e4547f6c5
BUG: 1184344
Signed-off-by: Avra Sengupta <asengupt@redhat.com>
Reviewed-on: http://review.gluster.org/9470
Reviewed-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Kaushal M <kaushal@redhat.com>
Also corrected few things related to rdma documentation
in other docs.
Change-Id: If35d7e2e884974ff3719f6a9f111f7b45ac8c0ec
Signed-off-by: Raghavendra Talur <rtalur@redhat.com>
Reviewed-on: http://review.gluster.org/9443
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: jiffin tony Thottan <jthottan@redhat.com>
Reviewed-by: Ben England <bengland@redhat.com>
Reviewed-by: Humble Devassy Chirammal <humble.devassy@gmail.com>
Problem:
When all the bricks are down at the time of mounting the volume, then mount
command hangs.
Fix:
1. Ignore all CHILD_CONNECTING events comming from subvolumes.
2. On timer expiration (without enough up or down childs) send
CHILD_DOWN.
3. Once enough up or down subvolumes are detected, send the appropriate event.
When rest of the subvols go up/down without changing the overall
ec-up/ec-down send CHILD_MODIFIED to parent subvols.
Change-Id: Ie0194dbadef2dce36ab5eb7beece84a6bf3c631c
BUG: 1179180
Signed-off-by: Pranith Kumar K <pkarampu@redhat.com>
Reviewed-on: http://review.gluster.org/9396
Reviewed-by: Xavier Hernandez <xhernandez@datalab.es>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
This patch solves some problems that caused dispersed volumes to not
pass posix smoke tests:
* Problems in open/create with O_WRONLY
Opening files with -w- permissions using O_WRONLY returned an EACCES
error because internally O_WRONLY was replaced with O_RDWR.
* Problems with entrylk on renames.
When source and destination were the same, ec tried to acquire
the same entrylk twice, causing a deadlock.
* Overwrite of a variable when reordering locks.
On a rename, if the second lock needed to be placed at the beggining
of the list, the 'lock' variable was overwritten and later its timer
was cancelled, cancelling the incorrect one.
* Handle O_TRUNC in open.
When O_TRUNC was received in an open call, it was blindly propagated
to child subvolumes. This caused a discrepancy between real file
size and the size stored into trusted.ec.size xattr. This has been
solved by removing O_TRUNC from open and later calling ftruncate.
Change-Id: I20c3d6e1c11be314be86879be54b728e01013798
BUG: 1161886
Signed-off-by: Xavier Hernandez <xhernandez@datalab.es>
Reviewed-on: http://review.gluster.org/9420
Reviewed-by: Dan Lambright <dlambrig@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Tested-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
gfapi consumers can invoke glusters_globals_init() multiple times
through glfs_new(). This will result in re-initialization of already
inited variables and non deterministic behavior. To avoid this, a
new function gf_globals_init_once() has been added. The invocation
of this function is guarded through pthread_once(), thereby ensuring
single initialization of per process globals.
Change-Id: I0ecde02ee49e0c7379c2eb0f1c879d89774ec82f
BUG: 1184366
Signed-off-by: Vijay Bellur <vbellur@redhat.com>
Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>
Reviewed-on: http://review.gluster.org/9430
Tested-by: Gluster Build System <jenkins@build.gluster.com>
* Bring in option to disable memory accounting for a glusterfs process
This reverses the changes done by the commit
7fba3a88f1ced610eca0c23516a1e720d75160cd.
* Change the key from "memory-accounting" to "no-memory-accounting", as by
default all the glusterfs process enable memory accounting now. So to
disable memory accounting for some process, "no-mem-accounting" argument has
to be passed.
Change-Id: I39c7cefb0fe764ea3e48f4e73e1305b084c5f497
BUG: 1184366
Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>
Reviewed-on: http://review.gluster.org/9469
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Currently the test case changed here, checks for the peer count to
be 1 until probe timeout and then checks for the changed
configuration, if it has been synced.
The peer count is not a gurantee that the configuration is also
in sync, hence changing this test case to check for the conf
update till probe timeout, by which time it should be in sync
(or at least that is our tolerance), and the test case deemed
as passing.
Change-Id: I4b1560979cfde3bd3bd691852d7d3a63e253bcf2
BUG: 1181203
Signed-off-by: Shyam <srangana@redhat.com>
Reviewed-on: http://review.gluster.org/9498
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>
Fix two spurious failures in tests/features/ssl-authz.t
1) Wait for bricks to come online after starting a volume, so that
the mount is usable without "socket not connected" error
2) For a mount that must fail, we may get the situation where there
is no mount at all, which means creating a file will write to the
mount point instead of failing. To cover that case, write the
file and check it is absent from the brick.
BUG: 1129939
Change-Id: If95e1d65ab23d11123f778c20f8110a3177b0e7f
Signed-off-by: Emmanuel Dreyfus <manu@netbsd.org>
Reviewed-on: http://review.gluster.org/9483
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Jeff Darcy <jdarcy@redhat.com>
problem : If quota is disabled then all the options associated with
quota is removed, except quota-deem-statfs and quota-timeout.
When gluster volume info is issued then the user can see that quota
is disabled whereas quota-deem-statfs and quota-timeout values still
exist.
Solution : remove quota-deem-statfs and quota-timeout option when quota is
disabled
NOTE : If features.quota-deem-statfs is turned on, it takes quota limits
into consideration while estimating fs size.
Change-Id: I8cca6a8f47d2355799228643aedc8fc03896cfad
BUG: 1151933
Signed-off-by: Sachin Pandit <spandit@redhat.com>
Reviewed-on: http://review.gluster.org/8924
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijaikumar Mallikarjuna <vmallika@redhat.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>
In case a new node is added to the peer, after a snapshot was
taken, the geo-rep files are not synced to that node. This
leads to the failure of snapshot restore. Hence, ignoring the
missing geo-rep files in the new node, and proceeding with
snapshot restore. Once the restore is successful, the missing
geo-rep files can be generated with "gluster volume geo-rep
<master-vol> <slave-vol> create push-pem force"
Change-Id: I1c364f8aefdd6c99b0b861b6d0cb33709ec39da2
BUG: 1181418
Signed-off-by: Avra Sengupta <asengupt@redhat.com>
Reviewed-on: http://review.gluster.org/9489
Reviewed-by: Sachin Pandit <spandit@redhat.com>
Reviewed-by: Aravinda VK <avishwan@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Krishnan Parthasarathi <kparthas@redhat.com>
If restore commit is successful on the originator and
a few nodes, but fails on some other node, restore cleanup
should restate the volume and the snapshot in question
as it was before the command was run.
Change-Id: I7bb0becc7f052f55bc818018bc84770944e76c80
BUG: 1181418
Signed-off-by: Avra Sengupta <asengupt@redhat.com>
Reviewed-on: http://review.gluster.org/9441
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Rajesh Joseph <rjoseph@redhat.com>
Reviewed-by: Atin Mukherjee <amukherj@redhat.com>
Reviewed-by: Kaushal M <kaushal@redhat.com>
* in readdirp callbak marker is calling inode_path on the inodes that
are not yet linked to the inode table.
Change-Id: I7f5db29c6a7e778272044f60f8e73c60574df3a9
BUG: 1176393
Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>
Reviewed-on: http://review.gluster.org/9320
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Tested-by: Raghavendra G <rgowdapp@redhat.com>
PROBLEM:
Gathering iatts from ONLY the hashed subvol during MKDIR and unwinding
them can cause md-cache to cache and serve these values for a while to
the application.
And then, at a later point of time, when a LOOKUP on either the dir or
its parent gathers attributes from all subvolumes of dht and things are
evened out as part of DHT_UPDATE_TIME, the application could be getting
a different set of [cm]times (i.e., one of the non-hashed subvolumes'
times could be selected by virtue of having the highest values),
causing it to think the directory underwent modification even when it
might not have.
The effect of this bug becomes apparent in programs like tar, which rely
on the ctime of the files before and after archiving a file to
ascertain that the file remained unchanged during this time.
FIX:
Aggregate iatts from ALL sub-volumes of DHT during MKDIR.
Change-Id: I04c4ca3e3b9552772e2b089be680f8afeb72089e
BUG: 1179169
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/9465
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Tested-by: Raghavendra G <rgowdapp@redhat.com>
That way, in the cbk, the latest values are updated in the cache.
Change-Id: Ia149e352e4763e8f5b910a3f4cb64d2dda4534b1
BUG: 1179169
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/9476
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Raghavendra G <rgowdapp@redhat.com>
Tested-by: Raghavendra G <rgowdapp@redhat.com>
Current Behaviour:
1. Geo-replication gsec_create creates common_secret.pem.pub file
containing public keys of the all the nodes of master cluster
in the location /var/lib/glusterd/
2. Geo-replication create push-pem copies the common_secret.pem.pub
to the same location on all the slave nodes with same name.
Problem:
Wrong public keys might get copied on to slave nodes in multiple
geo-replication sessions simultaneosly.
E.g.
A geo-rep session is established between Node1(vol1:Master) to
Node2 (vol2:Slave). And one more geo-rep session where
Node2 (vol3) becomes master to Node3 (vol4) as below.
Session1: Node1 (vol1) ---> Node2 (vol2)
Session2: Node2 (vol3) ---> Node3 (vol4)
If steps followed to create both geo-replication session is as
follows, wrong public keys are copied on to Node3 from Node2.
1. gsec_create is done on Node1 (vol1) -Session1
2. gsec_create is done on Node2 (vol3) -Session2
3. create push-pem is done Node1 - Session1.
-This overwrites common_secret.pem.pub in Node2
created by gsec_create in second step.
4. create push-pem on Node2 (vol3) copies overwrited
common_secret.pem.pub keys to Node3. -Session2
Consequence:
Session2 fails to start with Permission denied because of wrong
public keys
Solution:
On geo-rep create push-pem, don't copy common_secret.pem.pub
file with same name on to all slave nodes. Prefix master and
slave volume names to the filename.
NOTE: This brings change in manual steps to be followed to setup
non-root geo-replication (mountbroker). To copy ssh public
keys, extra two arguments needs to be followed.
set_geo_rep_pem_keys.sh <mountbroker_user> <master vol name> \
<slave vol name>
Path to set_geo_rep_pem_keys.sh:
Source Installation:
/usr/local/libexec/glusterfs/set_geo_rep_pem_keys.sh
Rpm Installatino:
/usr/libexec/glusterfs/set_geo_rep_pem_keys.sh
Change-Id: If38cd4e6f58d674d5fe2d93da15803c73b660c33
BUG: 1183229
Signed-off-by: Kotresh HR <khiremat@redhat.com>
Reviewed-on: http://review.gluster.org/9460
Reviewed-by: Aravinda VK <avishwan@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Venky Shankar <vshankar@redhat.com>
Tested-by: Venky Shankar <vshankar@redhat.com>
